MAP, real-world wins, and gen AI strategies
Industry leaders showed how to turn cloud migrations into measurable business gains by pairing AWS’s Migration Acceleration Program (MAP) with disciplined operating models and executive sponsorship. The Hartford and Adevinta detailed modernization at scale, from portfolio-wide R-factor choices to single-cloud consolidation, automation, and FinOps. With data platforms as the foundation, organizations are fast-tracking gen AI from pilots to production to unlock new value.
Points clés
- An IDC study found AWS migrations deliver, on average, a 31% reduction in total cost of ownership, a 62% boost in staff productivity, and 75% faster response to business needs
- AWS’s Migration Acceleration Program (MAP) structures cloud moves into three phases—assess, mobilize, migrate and modernize—guided by the seven R strategies from rehost to refactor and retire
- MAP spans 500+ global partners and includes financial investments to offset labor and the “double bubble” of running source and target environments
- The Hartford, a 214-year-old, $20B+ insurer aiming to double growth, is modernizing 400+ apps with intentional R-factor choices, scores for modernization, and low/no-touch DevSecOps pipelines
- The Hartford identified 30–40% of its portfolio not on existing roadmaps and is migrating it to cloud to avoid recurring on‑prem currency spend
- Adevinta operates 25 brands with 2.5B monthly visits, 5,000+ employees, and €1.8B in revenue; it acquired eBay Classifieds Group for $9B in 2021
- Post-acquisition, Adevinta pursued a single-cloud strategy on AWS, consolidating four data centers and ~20,000 servers, with the final two data centers slated to close next year
- Adevinta planned migration of ~3,600 services and tied outcomes to company-wide bonuses; lead time for change fell significantly during and after the 2023 AWS move
- McKinsey estimates gen AI could unlock $4.4T in incremental value; early adopters like BMW, DoorDash, and Thomson Reuters use Amazon Bedrock (and Amazon Connect) to operationalize gen AI
- Adevinta advanced 40 gen AI use cases to five in production (e.g., SEO automation, jobs CV/description generation) while solving GPU capacity constraints via AWS planning
À retenir
If you want cloud speed without the cloud hangover, start with a MAP, not a machete. Secure top-down sponsorship, pick your R-factors with intent, automate the boring bits (hello, DevSecOps), and teach everyone enough FinOps to keep the bill from becoming a horror story. Build your data platform early, pilot gen AI where the KPIs scream, and yes—tie some bonuses to outcomes; nothing motivates like a greener dashboard and a happier CFO.
Sources
Quiz sur la vidéo: 5 questions
